BF amazon-shopper
Amazon shopping that ends in a decision, not a page of links. It sweeps a dozen query phrasings instead of trusting one keyword, decodes the spec markings a listing hides (memory-card A2/V30 classes, active-ingredient concentration), and ranks on the metric that actually decides the buy — €/kg of active, €/GB, €/kg of protein — so the cheapest sticker price stops winning by default. Availability is a hard gate: a listing that still renders is not a listing you can buy. It reads amazon.es ANONYMOUSLY — there is no login, no cookie capture, no browser tab to share and no stored session, because 1.2.1 removed that code from the package rather than leaving it switched off. It reads no credential of your own; the only secrets it touches are API keys you set yourself for two optional paid paths (Apify, the Amazon Creators API), and with neither set it makes no third-party call. It searches amazon.es and nothing else — the classifieds scrapers earlier versions shipped behind a flag are gone. There is no write path to Amazon at all: it cannot add to a basket, place an order or change a setting. Built for the TinkerClaw fork — github.com/globalcaos/tinkerclaw. Use when the user asks "find me X on Amazon and tell me the best deal", "best price per kg/GB of X", or any iterative shopping conversation where the agent should drive the narrowing. NOT for: raw extraction without analysis, live price-tracking crons, delivery-date promises, or anything needing your account. See Permissions, Data Flow & Consent.
Amazon shopping that ends in a decision, not a page of links.
As a process F 46/100 · Will not run — References files that are not bundled: scripts/session-capture.mjs, scripts/amazon_fetch.py, scripts/relay-fetch.mjs
How to improve
- Shorten the description to 1024 characters.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 27. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1518 chars, limit 1024 - warning
body-longSKILL.md body ≈ 8633 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: scripts/session-capture.mjs - warning
missing-refreference to a missing file: scripts/amazon_fetch.py - warning
missing-refreference to a missing file: scripts/relay-fetch.mjs - note
description-budgetdescription takes 1518 of the ~15000-char shared budget for all skills - note
frontmatter-keyunknown frontmatter key "repository" - note
frontmatter-keyunknown frontmatter key "homepage"
Process rating: all ten parameters 46/100
- 0Tools and files. 3 referenced file(s) missing: scripts/session-capture.mjs, scripts/amazon_fetch.py, scripts/relay-fetch.mjs
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 30Running it twice. 35 mutating operations with no state check
- 40Consistency. Frontmatter name (amazon-shopper) differs from the folder (tinker-amazon-shopper)
- 40Execution cost. Instruction body is 8633 tokens: crowds the task out of the window
- 100Steps. 13 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 1 branches, has a failure section
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 18 top-level sections: this looks like several domains in one skill
- low The response is described with custom markup (5 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +3Description length 1518: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +5Description quotes 2 example trigger phrases
- +4Description says when NOT to use the skill
- +4Structure: 24 headings
- +3Step-by-step instructions: 13 items
- +4Has examples (10 code blocks)
- +3All 22 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 41.